An investigation on the Performance of iterative active Contour and atlas based mass Segmentation and classification of Mammographic and MR images

Abstract

Cancer is one of the leading causes of death among humans. There are newlinemany types of cancer which are found wide invariably. Among the various newlinetypes of cancer, breast cancer is the second leading of its types next to lung newlinecancer. A tumor can be benign or malignant. Masses are defined as lump that newlineis characterized by marginal properties. A benign is initial stage of cancer newlinewhere as malignant is latter of its part. After surveying the existing newlinetechniques, it is observed that there is a need for improvement in the newlinesegmentation of masses in mammogram and MR images to effectively newlineclassify the same. newlineFrom the literature surveyed it is observed that the discrimination of newlineforeground and background of the image is a challenging task. The accuracy newlineof the automatic segmentation algorithm is low. Many other segmentation newlinetechniques result in over segmentation. Hence the following approaches are newlineattempted in the current research to overcome the drawbacks of the existing newlinemethods. newlineIn the present research, three methods of segmentation are proposed, newlinenamely Iterative active contour mass segmentation of mammogram images, newlineAutomatic active contour segmentation of MR images and Atlas based newlinesegmentation of MR images. Classification is done applying iterative active newlinecontour method and Atlas based segmentation using Adaptive Neuro Fuzzy newlineInference System (ANFIS). newline